A surprise on Qobuz...

[voodooless]

I'd still be interested to see if anyone can make a better, more accurate one that could still be used to promote the service.

How about just being honest about it. This is what Apple says about lossless audio:

We developed our own implementation of AAC (Advanced Audio Codec) that delivers audio that’s virtually indistinguishable from the original studio recording. We also introduced Apple Digital Masters to deliver the highest-quality recordings. And now, we’re offering Apple Music subscribers the additional option to access our entire catalog encoded using lossless audio compression at no extra cost.

Zero mentions about the benefits of high-res audio.

 

[voodooless]

Newer DAC chips seems to have better filters that fully attenuate at 24KHz

Note that this is too late. Sampling rate for this test is 44.1 kHz. The filter should be done at 20.05 kHz at best

This seems to be some modern fad… I’m sure there is some rationale behind it.

 

[bluefuzz]

I can live with the slightly dirty feeling that I get knowing I'm subsidising moderate levels of misinformation but if it escalates it will be tidal for me when my subscription comes up.

I don't see Tidal's marketing copy as being much better. They still imply that HiRes is somehow 'better'. No stairsteps though. Here are a few selections from Tidal:

"Max – (Up to 24-bit, 192 kHz) – Experience best-in-class sound quality that opens up every detail with HiRes Free Lossless Audio Codec (HiRes FLAC)."

"High – (Up to 16-bit, 44.1 kHz) – Listen to over 100M songs in studio quality with FLAC. As an open source format, every artist can create and deliver high fidelity music with ease."

"The exact sound quality that artists and producers hear when they record music can be recreated through your speakers when listening in HiRes FLAC. To get the full experience, we suggest you research if the digital-to-analog converter (DAC) on your listening device can fully support the format (lossless 24-bit, 192 kHz) as well as use a wired or WiFi connection."

"HiRes FLAC is the distinction we make for any Free Lossless Audio Codec (FLAC) file that is greater than 16-bit, 44.1 kHz, which is the standard CD quality. HiRes FLAC on TIDAL can deliver superior sound in files as large as 24-bit, 192 kHz."

 

[IAtaman]

Note that this is too late. Sampling rate for this test is 44.1 kHz. The filter should be done at 20.05 kHz at best

This seems to be some modern fad… I’m sure there is some rationale behind it.

Yes, and not just ESS, AKM has the same/ similar setup. That's why I thought maybe using 48K might be better. Do you have access to any filter tests done at higher sampling rates by any chance?

 

[Haskil]

I think I was OK doing that, as you were responding to me in the context of a discussion about whether MQA files exist on Qobuz today!

As an aside, any streaming service that passes bit perfect lossless 24 bit files with all the information intact will "support" MQA files, so that tells us that Qobuz is passing tracks as supplied: a good thing. However, you need a DAC that does full processing of MQA files to know which is which.

We'll have to live with that, because if Qobuz was to implement anything that told us files were MQA they risk being sued by the patent troll that holds the MQA related DRM patent.

No, I responded very precisely by reminding a speaker who preferred Tidal to Qobuz that the latter provided PDFs of the record booklets when the publisher delivered them with the tracks and had refused to adopt MQA in place of PCM files full bandwidth when this codec was put forward and was adopted by many publishers... while Tidal had accepted these files encoded with loss. In this time, we thus found at Qobuz the same discs in PCM HR and at Tidal encoded with losses in MQA... Qobuz not wanting to broadcast this format...

Today Qobuz obviously does not offer MQA files without announcing them (your exemple) as such either in streaming or for sale... because the publisher would have only chosen this format... If he did so it would be "deception on the merchandise "and the first court that comes along would heavily condemn this platform...

The law is also very strict in France with the moral rights of performers and right of producer : a streaming platform is completely prohibited from modifying the files provided to it ; the only thing allowed is replay gain for streaming. So Qobuz cannot generate files itself...

 

[Haskil]

This also the case for the 24/192 2006 original of that release (https://open.qobuz.com/album/sxawmcgb6n9ha). It plays back as 24/44.1 and is transmitted at 44.1 but the DAC shows MQA/352.

Also, I tried some early releases which are marked as 16/44.1 CD-Quality and the DAC shows MQA/44.1. Example: https://open.qobuz.com/album/j82ajf4kvah5b

One release marked 24/48 is shown by the DAC as MQA/96. https://open.qobuz.com/album/kazufr7hlwhea

Also, some releases marked 24/44.1 are shown by the DAC as MQA/44.1. Example: https://open.qobuz.com/album/v9zkmzdl1d4wb

Since this release (https://open.qobuz.com/album/vaqrmh8i84cjc) is marked as 24/176 and shown by the DAC as non-MQA at 176, 2L clearly possess PCM-files of that format. This version is what is most likely on the pure audio blu ray disc since that allows for up to 24/192 PCM.

Several other recent releases that play back as MQA/352 are also available on pure audio blu ray and thus 24/176 PCM. Why were those not given to Qobuz instead of the MQA?

Curiously, "2L - the MQA experience" is available in DXD. How exactly does that have the MQA properties?

View attachment 347816
I use a MacBook Air M1 2020 running macOS 13.2.1 using the Qobuz app and having the output device set to the SMSL DL200.


Edit: Full volume in the app is required for MQA authentication and decoding. At lower volumes, the sample rate is shown as 44.1.

Thank you, it's very interesting : all these recordings come from the catalog of the same small publisher... who was the champion of MQA and who I remember, an amusing paradox, is one of those who are opposed to Dolby Atmos and fueled Stereophile's crusade against this lossy streaming format...

I'm going to do some listening: I don't have an MQA decoder but I can find out what format I'm sending into my DAC with Audirvana

PS.
I did some surveys... recordings of this editor advertised as HD 24/192 are actually played in 24/44.1 in streaming... without mentioning the fact that they are MQA... Audirvana tells me that.

Funny thing, when we go into the Qobuz app to read, it says : available in 24/192... but playback is in 24/44.1. So anyone who does not know that MQA exists will see if this format is available for purchase... they then click on the fatal link for the visa card... the recording is then only available in two formats: 24/96 and 16/44.1... And obviously never in multichannel even though the version exists... Where si 24/192 ???? For people who dont know anything of MQA il y a une grosse erreur...

 

[Haskil]

Sadly, they often don’t. For example, while they carry many Delos recordings (kudos!), none of them come with booklets, even though the vast majority of those are available for free to anyone from the Delos website. What’s worse, their metadata (even simple things like track names, artists and composers) are all over the place with Delos recordings.

The integrators who are responsible for providing streaming platforms and online sales sites for dematerialized files very often forget to provide the PDFs of the booklets for several reasons (they work poorly and so do the publishers) and above all provide Apple as a priority and Apple does not provide PDFs... so... very common forgetting...
The former boss of Qobuz was annoying them by asking these teams to check and request these booklets... Those days are over with the new boss and it's getting worse and worse... but at least when we have them, we have them with Qobuz and not the other major streaming players...


For the metadata: the only fault comes from the publishers who do anything... It turns out that I am a music critic and that I receive discs to download each time: a disaster... the metadata is limited to the naming of the file at the root... nothing else... Normal it's Wav... In certain cases, I have to reconstruct the works: I spent a day reconstituting the complete Mozart sonatas by Mao Fujita because all the sonatas had arrived in disorder without the title of the works being included in the naming of the files: only the indication of the name of the movement... and allegros, andante, there are many in these sonatas. ..
Only a few small publishers do things more or less correctly and never really because they provide Wav or MP3 and nothing can be done to get them to go to Flac or ALAC... And nothing helps.

Which means that on all streaming platforms we can listen to records where we don't know who is playing track by track... This is THE major problem with online music... and idiophiles prefer to compare the 16 /44.1 to 24/96 or 192...and decree that it's better than the CD format...like some people discussed the sex of angels while the world collapsed around them

 

[voodooless]

Yes, and not just ESS, AKM has the same/ similar setup. That's why I thought maybe using 48K might be better. Do you have access to any filter tests done at higher sampling rates by any chance?

They should be the same, just scaled to that rate.

The logic in this can be seen here, courtesy of Archimago:

Here you see that you do not expect to have any mirror images directly after half the sample rate. Therefore there is kind of a save area where you can actually extend the filter a bit and not have it cause any major issues.

 

[Galliardist]

No, I responded very precisely by reminding a speaker who preferred Tidal to Qobuz that the latter provided PDFs of the record booklets when the publisher delivered them with the tracks and had refused to adopt MQA in place of PCM files full bandwidth when this codec was put forward and was adopted by many publishers... while Tidal had accepted these files encoded with loss. In this time, we thus found at Qobuz the same discs in PCM HR and at Tidal encoded with losses in MQA... Qobuz not wanting to broadcast this format...

Today Qobuz obviously does not offer MQA files without announcing them (your exemple) as such either in streaming or for sale... because the publisher would have only chosen this format... If he did so it would be "deception on the merchandise "and the first court that comes along would heavily condemn this platform...

The law is also very strict in France with the moral rights of performers and right of producer : a streaming platform is completely prohibited from modifying the files provided to it ; the only thing allowed is replay gain for streaming. So Qobuz cannot generate files itself...

You responded, precisely, to me - see post 64 of this thread. It's not worth arguing about though, as you also did respond to someone else about the PDF booklets. It's all useful.

To be honest, even when we get things wrong, it's still useful as long as we get to the right conclusion.

I wasn't aware about the French law in this area, but it gets interesting with the MQA stuff. The example I posted is marked MQA because that version was specifically an MQA remaster. But can you tell me where the ones confirmed by JIW in post 95 here are marked as MQA in the Qobuz Windows app or the web links, because I certainly can't find it. There is an MQA CD logo in some of the booklets, but that wouldn't apply to the 24 bit quality streams, would it?

Sorry to sound antagonistic. I hope it is me missing something.

 

[Multicore]

Personally I have no use for high rez. I stream Spotify to my BT earbuds, and output analog from the cheapo DAC in the WiiM mini right into my Genelecs.

And I bet it sounds great.

 

[Dumdum]

I have less of a problem with the fact Qobuz is using a heaping helping of healthy horseshite to promote its hi-res offering than I have with its unavailability on platforms like NVidia shield, and Amazon Kindle. I want a streaming service that gives CD quality on all my platforms (Windows PC, NVidia Shield, Android Phone, Amazon Tablet), and that does it with native apps. Tidal does that Qobuz does not. That's why I'm now Team Tidal instead of Team Qobuz.

Tidal offers cd quality 44.1khz on android devices? New one on me, they did when they offered exclusive mode (have you not updated the app since then?) using an external dac, but you’re listening to resembled 48khz via the android audio layer or using a third party app to achieve it… so hardly as magical as you imply

And Amazon tablets resample also

I’d suggest checking actual sample rates before making big statements like you just did

 

[Multicore]

Absolutely. That was the first thing I noticed too. There's no doubt Qobuz' graphic is a mess but I'd still be interested to see if anyone can make a better, more accurate one that could still be used to promote the service.

It's doable. I think we've already seen here better presentations of the differences of sample rate and PCM bit depth. That should be separated from comparing PCM and MP3. These are important technical differences that can and should be explained to users.

Or is the consensus such that anyone selling Hi-Res should be tarred and feathered as a common criminal? I was of the impression that Qobuz is a fairly popular service among the ASR crowd, but it seems not ...

It's debatable here at ASR. Some are keen on Hi-Res and others not. Some see potential there and other think it has little or nothing to offer. In the extremes some seek equipment with better than 22-bit clean performance and others argue that Hi-Res is another kind of audiopoolery scam.

Check this out, one of the slides Amir presented at the Audio Engineering Society 2023. I think it is very interesting.

When we are discussing the relative merits of DACs in the blue and green zones, those are all differences that can only be practically realized in audio playback with Hi-Res program material. AMAZING!

When Amir uses his AP to reveal the properties of a new entrant in the top ranks of the Blue, we're using a microscope to judge the quality of the smoothness of the boundary of the ink in the curves of the serifs of the print of the font on the page of a book.

This suggests ASR has many admirers of Hi-Res audio.

Where is the boundary of human audio perception? I think that's a tricky question to answer. Characterizing it in a few numbers can be useful, clearly. However, if we take a bandwidth and SNR spec that has some consensus as a rule of thumb in engineering and use it as an assumption in further mathematical reasoning, as we sometimes see here, for example Monty Montgomery (the best Youtube Splainer ever), we can make a map-territory error. Who knows that the band limits are for you or me or the next person? Are they stable? Can they be changed? Are band limits the right way to characterize my hearing or are they an emergent feature of the history of test equipment and protocols?

I am the pragmatic antiaudiophile. I'm repelled by audiophile aesthetics and lifestyle but I enjoy good sound. We've got a power amp that's got way more headroom than I will ever use because I don't know for sure how much I will need and it wasn't very expensive. That represents the kind of pragmatism I'm into. The DAC/preamp is similar. It's in Amir's Blue zone so it's transparent, safely in the forget-about-it category. That's a pragmatic psychic win, like the over powered amp.

Hi-Res ought to be the same but there's a practical problem: the fuss of getting good source material. The specs of the stream or file mean very little. To be someone who cares about Hi-Res I would need to check the files forensically to know what's what and develop trust in Hi-Res labels/distrubutors. I can't be bothered with that so I sweep it aside. CDs sound great. Good enough. See, again, a pragmatic trick that works for me to simplify my enjoyment of music.

But that's just me. Sometimes I like to imagine that Hi-Res is like silver audiophile USB cables, i.e. something to sell to people who know something but not all the relevant things, because the thought amuses me. (I'm as prone to pretentions of intellectual and aesthetic superiority as the next nerd.) But if I'm honest, Idk, maybe there are people who can experience improvments over 16 bit PCM with certain recordings in ideal listening conditions. Too much trouble for me but lots of people have fussy, expensive hobbies so why not?

 

[Count Arthur]

Tidal offers cd quality 44.1khz on android devices? New one on me, they did when they offered exclusive mode (have you not updated the app since then?) using an external dac, but you’re listening to resembled 48khz via the android audio layer or using a third party app to achieve it… so hardly as magical as you imply

And Amazon tablets resample also

I’d suggest checking actual sample rates before making big statements like you just did

Tidal is now lossless on Android:

 

[Multicore]

I don't see Tidal's marketing copy as being much better. They still imply that HiRes is somehow 'better'. No stairsteps though. Here are a few selections from Tidal:

"Max – (Up to 24-bit, 192 kHz) – Experience best-in-class sound quality that opens up every detail with HiRes Free Lossless Audio Codec (HiRes FLAC)."

"High – (Up to 16-bit, 44.1 kHz) – Listen to over 100M songs in studio quality with FLAC. As an open source format, every artist can create and deliver high fidelity music with ease."

"The exact sound quality that artists and producers hear when they record music can be recreated through your speakers when listening in HiRes FLAC. To get the full experience, we suggest you research if the digital-to-analog converter (DAC) on your listening device can fully support the format (lossless 24-bit, 192 kHz) as well as use a wired or WiFi connection."

"HiRes FLAC is the distinction we make for any Free Lossless Audio Codec (FLAC) file that is greater than 16-bit, 44.1 kHz, which is the standard CD quality. HiRes FLAC on TIDAL can deliver superior sound in files as large as 24-bit, 192 kHz."

That doesn't bother me as much except for "The exact sound quality that artists and producers hear when they record music can be recreated through your speakers when listening in HiRes FLAC." which is an absurd statement on at least two levels.

We can and should separate technical education of potential customers from arguments about the relative values of formats. The former should be correct. The latter will always be highly debatable because it involves questions like: Value to whom? For what purpose? In what context? and many more.

 

[voodooless]

That doesn't bother me as much except for "The exact sound quality that artists and producers hear when they record music can be recreated through your speakers when listening in HiRes FLAC." which is an absurd statement on at least two levels.

Maybe they put the impulse response of the speakers and room in the FLAC metadata, so you can convoluted your own speakers and room to sound like the studio it was recorded in

 

[Multicore]

Maybe they put the impulse response of the speakers and room in the FLAC metadata, so you can convoluted your own speakers and room to sound like the studio it was recorded in

And some magic to undo the audio processing after the artists and producers recorded the music.

 

[JIW]

You responded, precisely, to me - see post 64 of this thread. It's not worth arguing about though, as you also did respond to someone else about the PDF booklets. It's all useful.

To be honest, even when we get things wrong, it's still useful as long as we get to the right conclusion.

I wasn't aware about the French law in this area, but it gets interesting with the MQA stuff. The example I posted is marked MQA because that version was specifically an MQA remaster. But can you tell me where the ones confirmed by JIW in post 95 here are marked as MQA in the Qobuz Windows app or the web links, because I certainly can't find it. There is an MQA CD logo in some of the booklets, but that wouldn't apply to the 24 bit quality streams, would it?

Sorry to sound antagonistic. I hope it is me missing something.

I just checked and even selecting 16/44.1 CD-quality, the 2L tracks play back as MQA and Qobuz even shows them playing at 24/44.1. It seems 2L just provided one file for each release which in almost all cases probably is MQA.

Even the 24/176 PCM release only plays at that format.

 

[Haskil]

Vous m'avez répondu précisément - voir le message 64 de ce fil. Cela ne vaut pas la peine d'en discuter, car vous avez également répondu à quelqu'un d'autre à propos des livrets PDF. Tout est utile.

Pour être honnête, même lorsque nous nous trompons, cela reste utile tant que nous arrivons à la bonne conclusion.

Je ne connaissais pas la loi française en la matière, mais ça devient intéressant avec les trucs MQA. L'exemple que j'ai publié est marqué MQA car cette version était spécifiquement un remaster MQA. Mais pouvez-vous me dire où ceux confirmés par JIW dans le post 95 ici sont marqués comme MQA dans l'application Windows Qobuz ou dans les liens web, car je ne les trouve certainement pas. Il y a un logo de CD MQA dans certains livrets, mais cela ne s'appliquerait pas aux flux de qualité 24 bits, n'est-ce pas ?

Désolé de paraître hostile. J'espère que c'est moi qui manque quelque chose.

I did a long search in the archives of this editor present on Qobuz, following the very interesting message from another participant in these discussions and informed him in my response of my comments and thanked him for it.

Concerning these L2 discs offered in streaming and for sale on Qobuz: in streaming, there are clearly MQAs not indicated as such given the format change made during playback (24/192 becomes 24/44.1). But for sale, the same recordings are in PCM not encoded in MQA... (choice of 24/96 or 16/44.1 and not in multichannel when the SACD version offers this layer)...

But these are the discs of a very particular small publisher which made the ostensible choice of MQA (and which ironically spits on Dolby Atmos because it is lossy compressed). Because to listen to lots of new releases every day on Qobuz (it's my job to do so), I have never encountered MQA on a Warner, Decca, Harmonia Mundi, DG, RCA, Sony, Mirare, Erato, Alpha, etc. If you have one or more to report to me, I am very interested in one of my articles and to annoy Qobuz.
But something tells me that they don't even know what they have put online in detail... because I have pointed out to them several times errors that they are struggling to correct... But I am going to point out these to them MQA tracks not reported... (well, they don't sell them!)

 

[Dumdum]

Tidal is now lossless on Android:

I’m always a bit dubious about tidal, they may well have uptaken android 14’s bit perfect ability now, but it won’t be on android 13 or prior… so I’d take the time to check what is happening

 

[theREALdotnet]

It's straight up lying, not that the 'hires' moniker is a technical standard anyway, it's marketing bullshit that's not worth the pixels it's written on from a technical standpoint.

It‘s clear that the term hi-res was born out of a misunderstanding of the technology by the people who coined it. It invokes the coarse vs fine stair steps simile that seems so hard to get rid of. If anything, formats using >44.1kHz or >16bit samples should be called hi-bandwidth or hi-SNR, respectively. That of course would put the spotlight on the actual SNR and bandwidth of the material, which would then likely reveal that the choice of digital format isn’t the bottleneck here.

I personally would choose formats like 96/24 over 44.1/16, simply because I like the added SNR headroom for digital processing on the playback side. After volume levelling and digital volume control, DSP or Dirac I can easily be -20 to -30dB down from the original signal level.

I think the higher bandwidth of 96kHz can be beneficial for the selection of less intrusive anti-aliasing filters, but at my age I don’t care about anything above 16kHz, so 44.1 is fine for me. It’s just that 44.1/24 material is not so common.